July 30, 2025
5 min read
Anirban Ghoshal
Google's new toolset enables AI agents to query BigQuery data, enhancing agentic applications with enterprise context and accuracy.
Google has introduced a new toolset designed to help enterprises connect their AI agents directly to data stored in BigQuery, addressing the growing demand for agentic applications. Agentic applications are AI-powered systems capable of performing tasks autonomously without manual intervention. These applications are increasingly popular among enterprises as they enable more efficient use of limited resources.
One challenge enterprises face is providing AI agents with sufficient context to generate accurate responses to user requests. Google’s new toolset aims to solve this by allowing AI agents to execute queries and retrieve metadata from BigQuery, a cloud-based data warehouse.
The toolset includes several key tools:
- listdatasetids: Retrieves all dataset IDs within a Google Cloud project.
- getdatasetinfo: Provides detailed metadata about a specific dataset.
- listtableids: Lists all table IDs within a dataset.
- gettableinfo: Fetches metadata for individual tables.
- execute_sql: Allows running SQL queries directly in BigQuery and retrieving results. However, this toolset cannot be used standalone. Enterprises must implement it alongside Google’s open-source Agent Development Kit (ADK) and MCP Toolbox for Databases (formerly known as Generative AI Toolbox for Databases) to connect AI agents to BigQuery effectively. To use the ADK, enterprises assign the toolset to an agent created within the framework. This is done by importing the toolset from the
- How to use Google Gemini for Smarter Crypto Trading
- AI-Driven Crypto Trading Tools Reshape Market Strategies in 2025
- AI Agents Capabilities, Risks, and Growing Role
- AI Crypto Coins Drive 2025 Innovation as Blockchain and AI Converge
agents.tools module within a Python environment using the ADK command line interface (CLI) and software development kit (SDK). Additionally, enterprises can use the tool_filter parameter to control which tools are exposed to the agent.
The MCP Toolbox for Databases natively supports BigQuery’s pre-built toolset. Enterprises need a Python-supported environment where they create a new mcp-toolbox folder alongside their ADK-developed agentic application and install the MCP Toolbox. Google also allows defining custom SQL tools within the MCP Toolbox deployment.
Industry experts recognize the significance of this integration. Forrester vice president and principal analyst Charlie Dai noted that Google’s ADK and MCP integration reduces development overhead by eliminating custom integration work and enables AI agents to leverage enterprise data context for more accurate responses.
Google is not alone in this space. Competitors like Databricks, Snowflake, and Teradata have recently introduced MCP Servers and related offerings to help enterprises connect AI agents to their data lakehouses and databases.
While Google plans to expand this toolset with additional tools, no timeline for these updates has been provided.
Source: Google’s new toolset to help connect AI agents to BigQuery by Anirban Ghoshal, InfoWorld, July 30, 2025.
Frequently Asked Questions (FAQ)
Google's BigQuery Toolset for AI Agents
Q: What is the primary purpose of Google's new toolset for BigQuery? A: The primary purpose is to enable enterprises to connect their AI agents directly to data stored in BigQuery, facilitating the development of agentic applications by providing AI agents with the necessary context for accurate responses. Q: What are agentic applications? A: Agentic applications are AI-powered systems designed to perform tasks autonomously without requiring manual intervention from users. Q: Which tools are included in Google's new toolset for BigQuery? A: The toolset includeslist<em>dataset</em>ids, get<em>dataset</em>info, list<em>table</em>ids, get<em>table</em>info, and execute_sql.
Q: Can Google's new BigQuery toolset be used independently?
A: No, the toolset requires implementation alongside Google's open-source Agent Development Kit (ADK) and the MCP Toolbox for Databases.
Q: How are the tools integrated with AI agents using the ADK?
A: The tools are assigned to an agent within the ADK framework by importing them from the agents.tools module in a Python environment. The tool_filter parameter can be used to control which tools are exposed to the agent.
Q: What role does the MCP Toolbox for Databases play?
A: The MCP Toolbox for Databases natively supports BigQuery's pre-built toolset and requires a Python-supported environment for its deployment alongside ADK-developed agentic applications. Custom SQL tools can also be defined within the MCP Toolbox.
Q: What are the benefits of this integration according to industry experts?
A: According to Forrester VP Charlie Dai, the integration reduces development overhead by eliminating custom integration work and allows AI agents to leverage enterprise data context for more accurate responses.
Q: Are there competitors offering similar solutions?
A: Yes, competitors such as Databricks, Snowflake, and Teradata have introduced similar offerings, like MCP Servers, to connect AI agents with their data lakehouses and databases.
Q: Will Google expand this toolset further?
A: Google plans to expand the toolset with additional tools, though a specific timeline for these updates has not yet been provided.